Probe placement detection methods, systems, and procedures for otoacoustic emission testing

Abstract

This invention belongs to the field of medical testing technology, specifically disclosing a probe placement detection method, system, and program product for otoacoustic emission (OAE) testing. By constructing a multi-dimensional comprehensive algorithm framework, it can simultaneously utilize low-frequency leakage characteristics, stimulus consistency characteristics, noise isolation characteristics, and stability characteristics for multi-dimensional feature joint judgment and scoring to quantify the probe's airtightness. Furthermore, it adopts a dual-track fusion mechanism of using scoring rules and machine learning to balance interpretability and robustness. This invention can automatically complete the probe airtightness determination without relying on human experience; through the comprehensive judgment of multi-dimensional features, it can avoid misjudgment based on a single indicator; it can reduce the number of invalid measurements, saving testing time; and it effectively ensures the stability of otoacoustic emission measurement signals, improving the clinical reliability of test data.

Description

Technical Field

[0001] This invention belongs to the field of medical testing technology, specifically relating to a probe placement testing method, system, and program product for otoacoustic emission testing. Background Technology

[0002] Otoacoustic emissions (OAE) testing is one of the fundamental methods in clinical hearing examination. Its principle involves using a speaker on a probe to emit specific sound signals, which are conducted through the outer, middle, and inner ear to induce responses in the outer hair cells of the cochlea. The microphone on the probe collects and analyzes the induced sounds from these hair cells to determine if their function is normal. During OAE testing, it is crucial to ensure the probe is correctly placed within the ear canal, with its tip not obstructed by the canal wall, and to create a good acoustic seal; otherwise, inaccurate testing or misdiagnosis may occur.

[0003] Otoacoustic emission (OAE) testing often relies on the user's experience to determine earplug suitability, lacking an objective and automated assessment mechanism, thus impacting testing efficiency and accuracy. While some existing devices offer simple probe level calibration, they cannot comprehensively assess "sealing," "external noise shielding effectiveness," and "probe stability," therefore, issues of misjudgment and low efficiency persist. Summary of the Invention

[0004] The purpose of this invention is to provide a probe placement detection method, system, and program product for otoacoustic emission testing, in order to solve the above-mentioned problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] Firstly, a probe placement detection method for otoacoustic emission testing is provided, including:

[0007] When performing otoacoustic emission testing using a probe, the ear canal response signal collected by the probe microphone inside the ear canal and the noise signal collected by the reference microphone are acquired. Both the probe microphone inside the ear canal and the reference microphone are mounted on the probe. The probe microphone inside the ear canal is used to collect the sound signal inside the ear canal to obtain the ear canal response signal, and the reference microphone is used to collect the environmental noise outside the ear canal to obtain the noise signal.

[0008] Low-frequency leakage features are extracted based on ear canal response signals to obtain low-frequency leakage features, and leakage scores are calculated using low-frequency leakage features.

[0009] Stimulus consistency features are extracted based on ear canal response signals to obtain stimulus consistency features, and stimulus scores are calculated using these features.

[0010] Noise isolation features are extracted based on ear canal response signals and noise signals to obtain noise isolation features, and noise scores are calculated using these noise isolation features.

[0011] Stability features are extracted based on ear canal response signals to obtain stability features, and stability scores are calculated using these features.

[0012] The leakage score, stimulus score, noise score, and stability score are weighted and calculated to obtain a comprehensive score.

[0013] The system determines whether the probe is correctly placed in the ear canal based on a comprehensive score and outputs a corresponding prompt signal.

[0014] In one possible design, the method further includes:

[0015] The low-frequency leakage feature, stimulus consistency feature, noise isolation feature and stability feature are used to form a fusion feature, and the fusion feature is input into the pre-trained Random Forest model for classification prediction to obtain the corresponding first prediction result;

[0016] The noise signal is processed by short-time Fourier transform to obtain a spectrum. The local time-frequency features of the spectrum are extracted by a one-dimensional convolutional neural network. The local time-frequency features are input into a pre-trained LSTM model for classification and prediction to obtain the corresponding second prediction result.

[0017] By combining the first prediction result and the second prediction result, a predicted classification result is obtained, which includes a closure score and a classification label.

[0018] Based on the prediction and classification results, determine whether the probe is correctly placed in the ear canal and output the corresponding prompt signal.

[0019] In one possible design, the method further includes:

[0020] The comprehensive score and the occlusion score are combined and calculated to obtain the final score. Based on the final score, it is determined whether the probe is correctly placed in the ear canal and the corresponding prompt signal is output.

[0021] In one possible design, the extraction of low-frequency leakage features based on the ear canal response signal to obtain low-frequency leakage features, and the calculation of a leakage score using the low-frequency leakage features, includes:

[0022] The ear canal response signal is processed by Fourier transform to extract the signal spectrum of the low-frequency part of the ear canal response signal, and the average sound pressure value of different frequency bands in the signal spectrum and the frequency slope of the signal spectrum are determined.

[0023] The average sound pressure level of different frequency bands is compared with the average target sound pressure level of the corresponding frequency band to determine the difference in sound pressure level of different frequency bands, and the average difference in sound pressure level of each frequency band is calculated.

[0024] The average difference in sound pressure levels across each frequency band and the frequency slope are substituted into a preset leakage scoring formula to calculate the leakage score. The leakage scoring formula is as follows:

[0025] LeakScore=100-clamp[a1×mean|ΔL|+b1×max(0,S 6), 0, 100]

[0026] Wherein, LeakScore represents the leakage score, clamp represents the clamp function, mean|ΔL| represents the mean difference in sound pressure level across frequency bands, S represents the frequency slope, and a1 and b1 are the first and second weighting coefficients, respectively.

[0027] In one possible design, the extraction of stimulus consistency features based on ear canal response signals to obtain stimulus consistency features, and the calculation of stimulus scores using these features, includes:

[0028] Determine the target level deviation between the signal level of the ear canal response signal and the preset target level;

[0029] Short-term jitter is determined based on the target level deviation obtained from the current measurement and the target level deviation obtained from the previous measurement.

[0030] The target level deviation and short-term jitter measured at the moment are substituted into a preset stimulus scoring formula for calculation to obtain the stimulus score. The stimulus scoring formula is as follows:

[0031] StimulusScore=100 (a2×σ) L +b2×D L )

[0032] Where StimulusScore represents the stimulus score, σ L D represents the target level deviation. L This indicates short-term jitter, and a2 and b2 are the set third and fourth weighting coefficients, respectively.

[0033] In one possible design, the noise isolation feature extraction based on the ear canal response signal and the noise signal to obtain noise isolation features, and the calculation of a noise score using the noise isolation features, includes:

[0034] Determine the noise power spectral density of the noise signal in a set frequency band, and determine the coherence function between the ear canal response signal and the noise signal;

[0035] The noise power spectral density and coherence function are substituted into a preset noise scoring formula to obtain the noise score. The noise scoring formula is as follows:

[0036] NoiseScore=100 {a3×PSD+b3×Max[γ 2 (f)]}

[0037] Where NoiseScore represents the noise score, PSD represents the noise power spectral density, and γ 2 (f) represents the coherence function, Max represents taking the maximum value, and a3 and b3 are the set fifth and sixth weight coefficients, respectively.

[0038] In one possible design, the extraction of stability features based on the ear canal response signal to obtain stability features, and the calculation of a stability score using these stability features, includes:

[0039] Retrieve the currently measured ear canal response signal and the previously measured ear canal response signal, and determine the cross-correlation peak value and phase difference variance between the two measured ear canal response signals;

[0040] The stability score is obtained by substituting the cross-correlation peak value and phase difference variance into a preset stability scoring formula. The stability scoring formula is as follows:

[0041] StabilityScore=100 [a4×(1 ρ) + b4×σ Φ ]

[0042] Where StabilityScore represents the stability score, ρ represents the peak value of the cross-correlation, and σ Φ This represents the phase difference variance, and a4 and b4 are the set seventh and eighth weighting coefficients, respectively.

[0043] Secondly, a probe placement detection system for otoacoustic emission (OAE) testing is provided, comprising a signal acquisition unit, a multi-dimensional scoring unit, a comprehensive scoring unit, a model determination unit, and a comprehensive determination unit, wherein:

[0044] The signal acquisition unit is used to acquire the ear canal response signal collected by the probe microphone and the noise signal collected by the reference microphone when performing otoacoustic emission testing using a probe. The probe microphone and the reference microphone are both set on the probe. The probe microphone is used to collect the sound signal inside the ear canal to obtain the ear canal response signal, and the reference microphone is used to collect the environmental noise outside the ear canal to obtain the noise signal.

[0045] The multidimensional scoring unit is used to extract low-frequency leakage features based on the ear canal response signal, obtain low-frequency leakage features, and calculate leakage scores using low-frequency leakage features; extract stimulus consistency features based on the ear canal response signal, obtain stimulus consistency features, and calculate stimulus scores using stimulus consistency features; extract noise isolation features based on the ear canal response signal and noise signals, obtain noise isolation features, and calculate noise scores using noise isolation features; and extract stability features based on the ear canal response signal, obtain stability features, and calculate stability scores using stability features.

[0046] The comprehensive scoring unit is used to calculate a comprehensive score by weighting the leakage score, stimulus score, noise score, and stability score.

[0047] The model decision unit is used to construct a fusion feature by utilizing low-frequency leakage features, stimulus consistency features, noise isolation features, and stability features, and input the fusion feature into a pre-trained Random Forest model for classification prediction to obtain the corresponding first prediction result; it performs short-time Fourier transform processing on the noise signal to obtain a spectrogram, and uses a one-dimensional convolutional neural network to extract local time-frequency features from the spectrogram, inputting the local time-frequency features into a pre-trained LSTM model for classification prediction to obtain the corresponding second prediction result; it then fuses the first and second prediction results to obtain the predicted classification result.

[0048] The comprehensive judgment unit is used to perform comprehensive judgment calculation on the comprehensive score and the occlusion score to obtain the final score. Based on the final score, it determines whether the probe is correctly placed in the ear canal and outputs the corresponding prompt signal.

[0049] Thirdly, a probe placement detection system for otoacoustic emission testing is provided, including:

[0050] Memory, used to store instructions;

[0051] The processor is configured to read instructions stored in the memory and execute, according to the instructions, any one of the probe placement detection methods for otoacoustic emission testing described in the first aspect above.

[0052] Fourthly, a computer-readable storage medium is provided, on which instructions are stored, which, when executed on a computer, cause the computer to perform any one of the probe placement detection methods for otoacoustic emission testing described in the first aspect. A computer program product is also provided, which, when executed on a computer, performs any one of the probe placement detection methods for otoacoustic emission testing described in the first aspect.

[0053] Beneficial Effects: This invention constructs a multi-dimensional comprehensive algorithm framework that simultaneously utilizes low-frequency leakage characteristics, stimulus consistency characteristics, noise isolation characteristics, and stability characteristics for joint judgment and scoring of multi-dimensional features to quantify the occlusion of probe examinations. Furthermore, it employs a dual-track fusion mechanism combining scoring rules and machine learning, balancing interpretability and robustness. This invention can automatically determine the occlusion of probe examinations without relying on human experience; through comprehensive judgment of multi-dimensional features, it avoids misjudgment based on a single indicator; it reduces the number of invalid measurements, saving testing time; and it effectively ensures the stability of otoacoustic emission measurement signals, improving the clinical reliability of test data. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart illustrating the method in Embodiment 1 of the present invention;

[0056] Figure 2 This is a schematic diagram of the system configuration in Embodiment 2 of the present invention;

[0057] Figure 3 This is a schematic diagram of the system configuration in Embodiment 3 of the present invention. Detailed Implementation

[0058] It should be noted that the descriptions of these embodiments are intended to aid in understanding the invention and do not constitute a limitation thereof. The specific structural and functional details disclosed herein are merely for describing exemplary embodiments of the invention. However, the invention may be embodied in many alternative forms and should not be construed as being limited to the embodiments described herein.

[0059] It should be understood that, unless otherwise explicitly specified and limited, the corresponding terms should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments according to the specific circumstances.

[0060] Specific details are provided in the following description to provide a complete understanding of the exemplary embodiments. However, those skilled in the art will understand that the exemplary embodiments can be implemented without these specific details. For example, the system may be shown in block diagrams to avoid obscuring the example with unnecessary details. In other embodiments, well-known processes, structures, and techniques may be shown without non-essential details to avoid obscuring the embodiments.

[0061] Example 1:

[0062] This embodiment provides a probe placement detection method for otoacoustic emission (OAE) testing, which can be applied to corresponding OAE testing probe processors. For example... Figure 1 As shown, the method includes the following steps:

[0063] S1. When performing otoacoustic emission testing using a probe, acquire the ear canal response signal collected by the probe microphone inside the ear canal, and the noise signal collected by the reference microphone.

[0064] In practice, when medical staff use the probe to perform otoacoustic emission (OAE) testing, the speaker on the probe emits a specific frequency sound signal (playing a low-frequency calibration tone). The microphone inside the ear canal collects the sound signal inside the ear canal, obtaining the ear canal response signal. The reference microphone on the probe is used to collect ambient noise outside the ear canal, obtaining the noise signal. The probe processor can acquire the ear canal response signal collected by the microphone inside the ear canal and the noise signal collected by the reference microphone in real time.

[0065] S2. Low-frequency leakage features are extracted based on the ear canal response signal to obtain low-frequency leakage features, and leakage scores are calculated using the low-frequency leakage features.

[0066] In practice, the principle of low-frequency leakage detection is that when the earplug is not sealed, a leakage channel is formed between the ear canal and the outside world, and the frequency response of the ear canal exhibits low-frequency high-pass characteristics. The features extracted by low-frequency leakage detection include the difference between the mean and target values ​​of low-frequency (250Hz–1000Hz) sound pressure levels (SPL) and low-frequency slope. The processor can perform Fourier transform processing on the ear canal response signal to extract the signal spectrum of the low-frequency portion (250Hz–1000Hz) of the ear canal response signal, and determine the average sound pressure level (SPL) of different frequency bands in the signal spectrum, as well as the frequency slope of the signal spectrum (if the slope is greater than 6dB, low-frequency leakage may exist); then, it compares the average SPL of different frequency bands with the average target SPL of the corresponding frequency band to determine the difference in SPL between different frequency bands (taking the absolute value of the difference), and calculates the average difference in SPL for each frequency band; then, it substitutes the average difference in SPL and the frequency slope into a preset leakage scoring formula to calculate the leakage score, which is:

[0067] LeakScore=100-clamp[a1×mean|ΔL|+b1×max(0,S 6), 0, 100]

[0068] Wherein, LeakScore represents the leakage score, clamp represents the clamp function, mean|ΔL| represents the mean difference in sound pressure level across frequency bands, S represents the frequency slope, and a1 and b1 are the first and second weighting coefficients, respectively.

[0069] S3. Stimulus consistency features are extracted based on ear canal response signals to obtain stimulus consistency features, and stimulus scores are calculated using stimulus consistency features.

[0070] In practical implementation, the principle of stimulus consistency detection is that otoacoustic emission measurement must ensure the stability of the stimulus level within the ear canal. If the measured level value differs too much from the target level value, or if the short-term jitter is too large, it indicates unstable coupling or that the probe is not properly positioned. The processor first determines the target level deviation between the signal level of the ear canal response signal and the preset target level; then, based on the target level deviation obtained from the current measurement and the target level deviation obtained from the previous measurement, it determines the short-term jitter; finally, it substitutes the currently measured target level deviation and the short-term jitter into a preset stimulus scoring formula to calculate the stimulus score. The stimulus scoring formula is as follows:

[0071] StimulusScore=100 (a2×σ) L +b2×D L )

[0072] Where StimulusScore represents the stimulus score, σ L D represents the target level deviation. L This indicates short-term jitter, and a2 and b2 are the set third and fourth weighting coefficients, respectively.

[0073] S4. Based on the ear canal response signal and noise signal, noise isolation features are extracted to obtain noise isolation features, and noise score is calculated using noise isolation features.

[0074] In practical implementation, the principle of noise isolation detection is that when the enclosure is good, the intensity of external noise coupled to the ear canal will decrease significantly. The processor first determines the noise power spectral density of the noise signal in a set frequency band and determines the coherence function between the ear canal response signal and the noise signal; then, it substitutes the noise power spectral density and the coherence function into a preset noise scoring formula to calculate the noise score, which is as follows:

[0075] NoiseScore=100 {a3×PSD+b3×Max[γ 2 (f)]}

[0076] Where NoiseScore represents the noise score, PSD represents the noise power spectral density, and γ 2 (f) represents the coherence function, Max represents taking the maximum value, and a3 and b3 are the set fifth and sixth weight coefficients, respectively.

[0077] S5. Based on the ear canal response signal, stability features are extracted to obtain stability features, and stability scores are calculated using stability features.

[0078] In practical implementation, the principle of stability detection is that if the probe position is unstable, the amplitude and phase of repeated measurements will be inconsistent. The specific method involves performing cross-correlation and phase consistency analysis on the responses of two adjacent measurements. The processor can retrieve the currently measured ear canal response signal and the previously measured ear canal response signal, and determine the cross-correlation peak value and phase difference variance between the two measured ear canal response signals; then, the cross-correlation peak value and phase difference variance are substituted into a preset stability scoring formula for calculation to obtain the stability score. The stability scoring formula is:

[0079] StabilityScore=100 [a4×(1 ρ) + b4×σ Φ ]

[0080] Where StabilityScore represents the stability score, ρ represents the peak value of the cross-correlation, and σ Φ The variance of the phase difference is represented by , and a4 and b4 are the set seventh and eighth weighting coefficients, respectively. The peak cross-correlation ρ represents the similarity between the two signals, and its calculation function is:

[0081]

[0082]

[0083] x(t) and y(t) represent the ear canal response signals from two consecutive measurements, respectively, with τ being the reference time. The calculation process for the phase difference variance includes: first, transforming the ear canal response signals from the two measurements into the frequency domain using Fourier transform to obtain the phase of the two measurement signals; then, calculating the phase difference Φ between the two measurement signals at each frequency point i. i Finally, the variance of the phase difference at each frequency point is calculated to obtain the phase difference variance σ. Φ :

[0084] .

[0085] S6. The leakage score, stimulus score, noise score, and stability score are weighted and calculated to obtain a comprehensive score.

[0086] In practice, different weights can be assigned to each detection item in advance, and then the leakage score, stimulus score, noise score, and stability score can be weighted and calculated to obtain a comprehensive score. For example, the formula for calculating the comprehensive score N is: N = LeakScore × 0.3 + StimulusScore × 0.2 + NoiseScore × 0.3 + StabilityScore × 0.2.

[0087] S7. Determine whether the probe is correctly placed in the ear canal based on the comprehensive score, and output the corresponding prompt signal.

[0088] In practice, the processor can determine whether the probe is correctly placed in the ear canal based on the comprehensive score N and output corresponding prompt signals. For example, if N≥80 and each individual score≥70, it can prompt "Probe is normal, test can be performed normally"; if 79≥N≥60, it can prompt "Probe needs fine adjustment"; if N<60, it can prompt "Probe needs to be repositioned or the earplug needs to be replaced".

[0089] This method can also employ a dual-track fusion strategy, using the comprehensive scoring rule as a safety net to ensure basic reliability. Simultaneously, it combines machine learning results for refined optimization, improving robustness in boundary conditions and complex scenarios. Specifically, the processor can utilize low-frequency leakage features, stimulus consistency features, noise isolation features, and stability features to form a fusion feature, which is then input into a pre-trained Random Forest model for classification prediction, yielding the corresponding first prediction result. The noise signal undergoes a short-time Fourier transform to obtain a spectrogram, and a one-dimensional convolutional neural network is used to extract local time-frequency features from the spectrogram. These local time-frequency features are then input into a pre-trained LSTM model for classification prediction, yielding the corresponding second prediction result. Finally, the first and second prediction results are fused to obtain a predicted classification result, which includes a seal score (SealScore) and a classification label (e.g., Good / Borderline / Fail; the classification label can also be used independently to determine whether the probe is correctly placed in the ear canal, so as to output the corresponding prompt signal). Finally, the system performs a comprehensive judgment calculation on the overall score and the occlusion score (such as weighted summation) to obtain the final score. Based on the final score, it determines whether the probe is correctly placed in the ear canal and outputs the corresponding prompt signal.

[0090] This method can automatically determine the tightness of probe examination without relying on human experience; through comprehensive judgment of multi-dimensional features, it can avoid misjudgment based on a single indicator; it can reduce the number of invalid measurements and save testing time; it can effectively ensure the stability of otoacoustic emission measurement signals and improve the clinical reliability of test data.

[0091] Example 2:

[0092] This embodiment provides a probe placement detection system for otoacoustic emission testing, such as... Figure 2 As shown, it includes a signal acquisition unit, a multi-dimensional scoring unit, a comprehensive scoring unit, a model determination unit, and a comprehensive determination unit, wherein:

[0093] The signal acquisition unit is used to acquire the ear canal response signal collected by the probe microphone and the noise signal collected by the reference microphone when performing otoacoustic emission testing using a probe. The probe microphone and the reference microphone are both set on the probe. The probe microphone is used to collect the sound signal inside the ear canal to obtain the ear canal response signal, and the reference microphone is used to collect the environmental noise outside the ear canal to obtain the noise signal.

[0094] The multidimensional scoring unit is used to extract low-frequency leakage features based on the ear canal response signal, obtain low-frequency leakage features, and calculate leakage scores using low-frequency leakage features; extract stimulus consistency features based on the ear canal response signal, obtain stimulus consistency features, and calculate stimulus scores using stimulus consistency features; extract noise isolation features based on the ear canal response signal and noise signals, obtain noise isolation features, and calculate noise scores using noise isolation features; and extract stability features based on the ear canal response signal, obtain stability features, and calculate stability scores using stability features.

[0095] The comprehensive scoring unit is used to calculate a comprehensive score by weighting the leakage score, stimulus score, noise score, and stability score.

[0096] The model decision unit is used to construct a fusion feature by utilizing low-frequency leakage features, stimulus consistency features, noise isolation features, and stability features, and input the fusion feature into a pre-trained Random Forest model for classification prediction to obtain the corresponding first prediction result; it performs short-time Fourier transform processing on the noise signal to obtain a spectrogram, and uses a one-dimensional convolutional neural network to extract local time-frequency features from the spectrogram, inputting the local time-frequency features into a pre-trained LSTM model for classification prediction to obtain the corresponding second prediction result; it then fuses the first and second prediction results to obtain the predicted classification result.

[0097] The comprehensive judgment unit is used to perform comprehensive judgment calculation on the comprehensive score and the occlusion score to obtain the final score. Based on the final score, it determines whether the probe is correctly placed in the ear canal and outputs the corresponding prompt signal.

[0098] Example 3:

[0099] This embodiment provides a probe placement detection system for otoacoustic emission testing, such as... Figure 3 As shown, at the hardware level, it includes:

[0100] The data interface is used to establish data communication between the processor and external data terminals;

[0101] Memory, used to store instructions;

[0102] The processor is configured to read instructions stored in the memory and execute the probe placement detection method for otoacoustic emission testing in Embodiment 1 according to the instructions.

[0103] Optionally, the system also includes an internal bus, through which the processor, memory, and data interface can be interconnected. This internal bus can be a PCIe (Peripheral Component Interconnect Eexpress) bus, which can be divided into an address bus, a data bus, a control bus, etc. The memory can include, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Flash Memory, First Input First Output (FIFO), and / or First In Last Out (FILO). The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0104] Example 4:

[0105] This embodiment provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the probe placement detection method for otoacoustic emission testing as described in Embodiment 1. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable systems.

[0106] This embodiment also provides a computer program product that, when run on a computer, executes the probe placement detection method for otoacoustic emission testing described in Embodiment 1. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable system.

[0107] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A probe placement and detection method for otoacoustic emission testing, characterized in that, include: When performing otoacoustic emission testing using a probe, the ear canal response signal collected by the probe microphone inside the ear canal and the noise signal collected by the reference microphone are acquired. Both the probe microphone inside the ear canal and the reference microphone are mounted on the probe. The probe microphone inside the ear canal is used to collect the sound signal inside the ear canal to obtain the ear canal response signal, and the reference microphone is used to collect the environmental noise outside the ear canal to obtain the noise signal. Low-frequency leakage features are extracted based on the ear canal response signal to obtain low-frequency leakage features, and leakage scores are calculated using these features, including: The ear canal response signal is processed by Fourier transform to extract the signal spectrum of the low-frequency part of the ear canal response signal, and the average sound pressure value of different frequency bands in the signal spectrum and the frequency slope of the signal spectrum are determined. The average sound pressure level of different frequency bands is compared with the average target sound pressure level of the corresponding frequency band to determine the difference in sound pressure level of different frequency bands, and the average difference in sound pressure level of each frequency band is calculated. The average difference in sound pressure levels across each frequency band and the frequency slope are substituted into a preset leakage scoring formula to calculate the leakage score. The leakage scoring formula is as follows: LeakScore=100-clamp[a1×mean|ΔL|+b1×max(0,S 6),0,100] Wherein, LeakScore represents the leakage score, clamp represents the clamp function, mean|ΔL| represents the mean difference in sound pressure levels across frequency bands, S represents the frequency slope, and a1 and b1 are the first and second weighting coefficients, respectively. Stimulus consistency features are extracted based on ear canal response signals to obtain stimulus consistency features, and stimulus scores are calculated using these features, including: Determine the target level deviation between the signal level of the ear canal response signal and the preset target level; Short-term jitter is determined based on the target level deviation obtained from the current measurement and the target level deviation obtained from the previous measurement. The target level deviation and short-term jitter measured at the moment are substituted into a preset stimulus scoring formula for calculation to obtain the stimulus score. The stimulus scoring formula is as follows: StimulusScore=100 (a2×σ L +b2×D L ) Where StimulusScore represents the stimulus score, σ L D represents the target level deviation. L This indicates short-term jitter, and a2 and b2 are the set third and fourth weighting coefficients, respectively. Noise isolation features are extracted based on ear canal response signals and noise signals to obtain noise isolation features, and noise scores are calculated using these features, including: Determine the noise power spectral density of the noise signal in a set frequency band, and determine the coherence function between the ear canal response signal and the noise signal; The noise power spectral density and coherence function are substituted into a preset noise scoring formula to obtain the noise score. The noise scoring formula is as follows: NoiseScore=100 {a3×PSD+b3×Max[γ 2 (f)]} Where NoiseScore represents the noise score, PSD represents the noise power spectral density, and γ 2 (f) represents the coherence function, Max represents taking the maximum value, and a3 and b3 are the set fifth and sixth weight coefficients, respectively; Stability features are extracted based on the ear canal response signal to obtain stability features, and a stability score is calculated using these features, including: Retrieve the currently measured ear canal response signal and the previously measured ear canal response signal, and determine the cross-correlation peak value and phase difference variance between the two measured ear canal response signals; The stability score is obtained by substituting the cross-correlation peak value and phase difference variance into a preset stability scoring formula. The stability scoring formula is as follows: StabilityScore=100 [a4×(1 ρ)+b4×σ Φ ] Where StabilityScore represents the stability score, ρ represents the peak value of the cross-correlation, and σ Φ This represents the phase difference variance, and a4 and b4 are the set seventh and eighth weighting coefficients, respectively. The leakage score, stimulus score, noise score, and stability score are weighted and calculated to obtain a comprehensive score. The system determines whether the probe is correctly placed in the ear canal based on a comprehensive score and outputs a corresponding prompt signal.

2. The probe placement and detection method for otoacoustic emission testing according to claim 1, characterized in that, The method further includes: The low-frequency leakage feature, stimulus consistency feature, noise isolation feature and stability feature are used to form a fusion feature, and the fusion feature is input into the pre-trained Random Forest model for classification prediction to obtain the corresponding first prediction result; The noise signal is processed by short-time Fourier transform to obtain a spectrum. The local time-frequency features of the spectrum are extracted by a one-dimensional convolutional neural network. The local time-frequency features are input into a pre-trained LSTM model for classification and prediction to obtain the corresponding second prediction result. By combining the first prediction result and the second prediction result, a predicted classification result is obtained, which includes a closure score and a classification label. Based on the prediction and classification results, determine whether the probe is correctly placed in the ear canal and output the corresponding prompt signal.

3. The probe placement detection method for otoacoustic emission testing according to claim 2, characterized in that, The method further includes: The comprehensive score and the occlusion score are combined and calculated to obtain the final score. Based on the final score, it is determined whether the probe is correctly placed in the ear canal and the corresponding prompt signal is output.

4. A probe placement detection system for otoacoustic emission testing, applied to the probe placement detection method for otoacoustic emission testing as described in any one of claims 1-3, characterized in that, It includes a signal acquisition unit, a multi-dimensional scoring unit, a comprehensive scoring unit, a model determination unit, and a comprehensive determination unit, wherein: The signal acquisition unit is used to acquire the ear canal response signal collected by the probe microphone and the noise signal collected by the reference microphone when performing otoacoustic emission testing using a probe. The probe microphone and the reference microphone are both set on the probe. The probe microphone is used to collect the sound signal inside the ear canal to obtain the ear canal response signal, and the reference microphone is used to collect the environmental noise outside the ear canal to obtain the noise signal. The multidimensional scoring unit is used to extract low-frequency leakage features based on the ear canal response signal, obtain low-frequency leakage features, and calculate leakage scores using low-frequency leakage features; extract stimulus consistency features based on the ear canal response signal, obtain stimulus consistency features, and calculate stimulus scores using stimulus consistency features; extract noise isolation features based on the ear canal response signal and noise signals, obtain noise isolation features, and calculate noise scores using noise isolation features; and extract stability features based on the ear canal response signal, obtain stability features, and calculate stability scores using stability features. The comprehensive scoring unit is used to calculate a comprehensive score by weighting the leakage score, stimulus score, noise score, and stability score. The model decision unit is used to construct a fusion feature by utilizing low-frequency leakage features, stimulus consistency features, noise isolation features, and stability features, and input the fusion feature into a pre-trained Random Forest model for classification prediction to obtain the corresponding first prediction result; it performs short-time Fourier transform processing on the noise signal to obtain a spectrogram, and uses a one-dimensional convolutional neural network to extract local time-frequency features from the spectrogram, inputting the local time-frequency features into a pre-trained LSTM model for classification prediction to obtain the corresponding second prediction result; it then fuses the first and second prediction results to obtain the predicted classification result. The comprehensive judgment unit is used to perform comprehensive judgment calculation on the comprehensive score and the occlusion score to obtain the final score. Based on the final score, it determines whether the probe is correctly placed in the ear canal and outputs the corresponding prompt signal.

5. A probe placement detection system for otoacoustic emission testing, characterized in that, include: Memory, used to store instructions; A processor is configured to read instructions stored in the memory and execute, according to the instructions, the probe placement detection method for otoacoustic emission testing as described in any one of claims 1-3.

6. A computer program product, characterized in that, When the computer program product is run on a computer, it performs the probe placement detection method for otoacoustic emission testing as described in any one of claims 1-3.

Images